Research for AI that works in the real world

Capable models are not enough. We research how to make AI systems reliable, controllable, and safe once they meet real data, tools, and people.

“A safe AI system is one where access, actions, and oversight are designed in from the start, not patched in later.”

Neobolt AI

Focus areas

Our research sits between powerful models and real deployment, where most AI systems succeed or fail.

Systems

How to build the software around a model so it can work with real data, tools, and workflows.

How should an AI agent break down tasks, call tools, and handle errors without derailing?

How do we ensure a model uses the right information for a task and cannot access what it should not?

How do models interact safely with APIs, databases, and enterprise systems that have their own state and rules?

Safety

How to limit what an AI system can do, catch failures before they cause harm, and keep people in control.

Designing access limits, permission boundaries, and audit logs into the architecture from the beginning.

How can a system recognize when an output is unreliable or dangerous before it is acted on?

When and how should a human be brought into the loop, and what information do they need to make a good decision?

Evaluation

How to measure whether an AI system is reliable on real tasks, not just on standard benchmarks.

Testing AI systems on the actual work they are meant to do, with realistic constraints and imperfect inputs.

Systematically mapping the ways systems break in practice, and designing tests that catch those specific failures.

How do we know a system is still behaving correctly after deployment, as data and conditions change?

Our early work is focused on foundations: how AI systems should be built, constrained, and tested.

Soon

Soon

Soon

We are an early team looking for people who want to work on hard, open problems in AI systems.